2026-05-15 — views $KLAC · KLA Corporation · Defect Inspection · Metrology · Reticle Inspection · Service · WFE — process control
KLA Corp (KLAC) — the process-control near-monopoly that compounds with every new node
KLAC owns ~80% market share at advanced-node defect inspection and metrology. The yield-loss math means fabs will pay almost anything to find defects early — and only one company sells the tool.
KLA Corporation (NASDAQ: KLAC) is the near-monopoly in semiconductor process control — the inspection and metrology tools that find defects, measure feature dimensions, and verify that what came out of the previous step is what was supposed to come out. At advanced nodes (5nm and below), KLA’s market share is roughly 80%. There is no single comparable position in any other WFE category. For builders, KLA is the cleanest “owns a moat” name in semiconductor equipment.
Four segments, two that matter
| Segment | What it does | Why it’s the moat |
|---|---|---|
| Defect inspection | Scans wafers to find physical defects (particles, scratches, pattern errors) | Broadband Plasma (BBP) line is the only credible advanced-node option |
| Metrology | Measures critical dimensions, film thickness, overlay accuracy | Optical and e-beam metrology — KLA’s lead is narrower but still dominant |
| Reticle inspection | Verifies the photomasks (reticles) are defect-free before printing | TerraSpec, Teron — protects against a single defective reticle ruining every wafer |
| Service | Recurring revenue on installed base | High-margin compounding revenue, structural baseline |
Defect inspection is where the moat is hardest. KLA’s Broadband Plasma (BBP) tools use a proprietary plasma light source plus optical detection to find sub-10nm defects on a 300mm wafer at production speed. The R&D investment required to develop a competitive BBP is enormous — Applied Materials has tried, others have tried, and none have meaningfully dented KLA’s advanced-node share. This is what an actual technology moat looks like.
The yield-loss math
Why does KLA command pricing power? Because of the math of yield loss at an advanced fab:
- One 300mm wafer at a 2nm node carries hundreds of dies worth ~$50K-$100K combined
- One undetected defect early in the process can ruin every die fabricated after it
- A single missed defect class affecting a wafer lot means $5M-$20M in lost product
- An inspection tool that catches the defect 5 steps earlier saves the entire downstream investment
Given those stakes, fabs will pay almost any price for a tool that finds the defect. KLA prices accordingly. Operating margins reflect this: KLA’s segment operating margin runs structurally higher than AMAT’s or Lam’s. The willingness-to-pay is what keeps competitors out — there’s no “low-cost KLA” segment to attack because the cost of a missed defect dwarfs the cost of the tool.
Why it compounds with every new node
Each new node has tighter tolerances. 5nm allows defect sizes that 3nm rejects. 3nm rejects what 2nm rejects more aggressively. Every node transition means:
- More inspection steps per wafer
- Tighter dimensional measurement requirements (KLA wins on metrology too)
- More reticle inspection iterations
- Higher KLA TAM per fab
This is rare: most semiconductor segments see unit price compression as nodes mature (etch per wafer roughly flat, deposition per wafer roughly flat). Process control is the exception — the dollar value of KLA equipment per wafer grows with each node transition because the inspection burden grows.
Why this matters for builders
For builders thinking about durability:
- Compute and memory are competitive layers. Multiple players, margin compression over time.
- WFE broad-portfolio (AMAT, LAM) is less competitive but still has multiple players in each category.
- Process control (KLA) is structurally a near-monopoly. This is the rarest competitive position in the entire AI supply chain.
The bull case is straightforward: more advanced nodes → more inspection per wafer → more KLA TAM. The bear case is that an ASML or Applied Materials breakthrough in metrology eventually erodes KLA’s lead, but no such breakthrough has appeared in 20+ years of trying.
Practitioner note
For builders evaluating exposure:
- KLA tracks fab investment, but with structurally higher growth. When AMAT WFE grows 1×, KLA tends to grow 1.2-1.5× because of the per-wafer TAM expansion described above. Read AMAT and Lam earnings as the rising-tide indicator; expect KLA to outperform on the way up.
- Operating margin is the durability signal. KLA’s operating margin staying stable through downcycles (when AMAT and Lam compress) is the marker that the moat is intact. If KLA’s margin starts compressing similarly to peers, that’s the first sign of competitive pressure.
- The reticle business is underappreciated. Reticle inspection grows with the number of unique chip designs — and AI-driven custom silicon (every hyperscaler with their own ASIC) means more reticles per fab. This is a structural growth lever that doesn’t get a lot of airtime in analyst models.
- China exposure is real but smaller than at AMAT or Lam. Process control is precisely the category US export controls target hardest (because it enables advanced-node manufacturing). The cap is binding, but advanced-node manufacturing outside China is what’s growing fastest anyway.
The under-considered angle: service is a 30%+ operating margin business that compounds with installed base. As KLA tools age, the service revenue compounds — and the installed base only grows. The next 10 years of service revenue is more visible than the next 10 years of any other line in this section.